Roughly a year ago, the creation of CRISPRed twins reignited fierce debate and trepidation about our oncoming era of designer babies. The experiment, designed to protect the babies against HIV, failed miserably. But He Jiankui’s brazen attempt against global ethical norms made it crystal clear that if the technology to engineer “better,” healthier offspring is available, some parents will flock to it.
But how far are we from being able to screen DNA for common “desirable” traits, such as higher IQ or height?
It’s difficult to talk about genetic selection without the umbrella of morality or ethics. Yet like all scientific advancements, it pays to ask where we are on the road down a troubling scientific journey—one we might not like the destination of. This month, a team led by Dr. Shai Carmi at the Hebrew University of Jerusalem asked just that. Given our current understanding of genetics, are we able to select for human embryos that turn into taller, smarter adults?
The answer is complicated, and your interpretation depends on your values. Using computer simulations, the team found that it is possible to increase the height and IQ of offspring by interrogating their DNA—a boost of roughly an inch in height and 2.5 points for IQ. These small increases come with significant caveats, in that the predictions often fail. In large nuclear families, for example, the majority of children whose DNA scored high for height are in reality not the tallest. The impact is so modest that scientists even question if it’s worth going through the trouble.
How this news lands for you also ultimately depends on where you’re at in the “treatment versus enhancement” debate, that is, if its ok to use genetic tools to screen for diseases or select for desirable traits. To Carmi, the results help bring a dose of reality to discussions about designer babies and embryo selection for reasons outside of serious medical conditions.
“We wanted to use these numbers or these methods to make the debate more quantitative, more evidence-based,” he said.
Others aren’t as convinced that the question of enhancement should even be asked. Just the idea is the “definition of eugenics” and “disturbing,” said genetics doctor Nicholas Katsanis at the Ann & Robert Lurie Children’s Hospital in Chicago, who was not involved in the study.
Regardless of which side of the debate you’re on, one thing is clear: we still have some time before we’re technologically capable of genetic selection (or medical eugenics), perhaps enough for more stringent regulations to be put into place.
“It may be desired to introduce oversight over at least the advertised outcomes,” said the team.
Similar to genetically altering embryos, genetic selection can be used—and often is—for good.
During in vitro fertilization (IVF), scientists routinely screen embryos for serious genetic defects before implantation, such as mutations that would render the embryo unviable or endow it with life-threatening disorders such as cystic fibrosis. This procedure, called preimplantation genetic diagnosis, is widely used and generally accepted as a screen for potential medical problems caused by a single gene.
As insight into human genetics exploded in the past decade, however, these screens have gotten increasingly complex. Scientists can now scan an entire genome based on a single cell, and our understanding of complex genetic traits is making it possible for a more detailed and thorough genetic diagnosis. It’s also how preimplantation genetic diagnosis is unwittingly entering a grey zone: is a small increase in risk for diabetes, heart disease, or Alzheimer’s sufficient to reject an embryo? And if so, why isn’t it acceptable—or even responsible—for parents to select for gene variants that protect their child from those disorders, giving them a higher chance of a healthy life?
Carmi is hardly naïve to the horrors of eugenics, a movement that justified atrocities such as forced sterilization and genocide. But his team decided to embrace that firestorm by first dealing with the underlying technology.
Complex traits often generate a “polygenic score,” or PS. Generally, the score is based on a population and not on an individual. But for PS to be practically used in embryo selection—ethical concerns aside—it needs to have sufficient predictive power to be able to differentiate between embryos, both within a single family and within a finite number of embryos. Although PS has been broadly used to examine complex traits and precision medicine in research, the authors said, so far no one’s really looked at the possibilities and limitations of that approach to embryo selection.
That’s what they set out to do, with the help of powerful computer simulations.
The team envisioned a hypothetical IVF cycle, during which the main goal was to identify embryos with two desired traits guided by multiple genes—increased height and IQ—compared to the average embryo. To find prospective parents, they used genetic data from longevity and schizophrenia studies, which contained profiles of at least some real-life spouses.
The team then artificially paired up other unrelated people in the test data set and used computer simulations to generate potential genetic profiles for the hypothetical offspring for each couple—real or otherwise. They took special care to ensure that the offspring profiles were as realistic as they could be, relying on a combination of parameters that critically matched genotype to phenotype—that is, the actual traits in a person.
Although an IVF cycle currently generates between three to eight viable embryos, the team simulated a wide number of offspring for each couple, placing a particular emphasis on 10. It’s likely that 10 represents the upper limit of how many embryos IVF can make even in the future, they explained.
For every embryo, the team used a simple genetic model to test for its potential height and IQ based on known literature. The results were surprisingly underwhelming: with five viable embryos, the average number of current IVF attempts, prospective parents can expect an increase of barely an inch in height and 2.5 points in IQ compared to the average embryo. Even with 10 embryos, a potential upper limit for future IVF attempts, the increase in height is just a mere three centimeters, or a smidge over an inch.
Not satisfied with a pure thought experiment, the team next compared their simulated results to actual observable data from 28 large families, each with 10 adult children on average. Each member had both their genomic data and phenotypical data available, making them a valuable dataset for a sanity check.
The results, again, were underwhelming. Using the same algorithm as their simulation, they found that the tallest child predicted by the computer was only the tallest in roughly a quarter of the families in real life. Even weirder, the predicted tallest kid was often shorter than the actual tallest sibling.
That’s not to say the algorithm needs dramatic improvement. More data could potentially boost its accuracy, but fundamentally complex traits like height and IQ aren’t all about data. Nutrition, lifestyle, socioeconomic status, and exposure to pollution, trauma, and yet unknown environmental factors (and genetic ones) could all contribute to how genes play out.
“There is much about these traits that is unpredictable,” said Carmi. “If someone selected an embryo that was predicted to have an IQ that was two points higher than the average, this is no guarantee it would actually result in that increase. There is a lot of variability that is not accounted for in the known gene variants.”
Although it’s possible that in the future complex genetic traits may receive a larger boost from selection, Carmi cautions that there’s still much we don’t know about human genetics to warrant using the technology—even if there were no ethical quandaries.
For one, the data was mostly obtained from people of European descent, meaning that the results may be even less applicable to populations from other parts of the world. Traits are also not cut and dry: higher IQ, something desirable, is linked to higher risk of autism and eating disorders. Parents eager to maximize more than one trait, if ever allowed, will run headfirst into tricky “designer” problems.
For all its controversy, Carmi’s study adds rigorous scientific data and shines a light on a complex topic. As Katsanis said, it is disturbing to contemplate the fundamental question of designer babies. But not contemplating it, before the next CRISPRed babies are born, is even more disturbing.